metadata
language:
- en
base_model:
- krea/Krea-2-Turbo
pipeline_tag: text-to-image
library_name: diffusers
widget:
- text: >-
A golden labrador running in a farm, while a pickup truck is driving
across a dusty road in the background far away. While a plane flies in the
sky. ethereal motion blur style
output:
url: images/00_turbo.png
- text: >-
A cat wearing glasses holding a stick to a board with the text 'He who
lives in glass houses shall not throw stones at others', while sitting on
a stool. ethereal motion blur style
output:
url: images/01_turbo.png
- text: >-
A man dressed in elegant knight armour and holding a sword pointing at the
sky while the ground around him cracks. ethereal motion blur style
output:
url: images/02_turbo.png
- text: >-
Two Nissan GT cars drifting by each other at a sharp corner in the road.
ethereal motion blur style
output:
url: images/03_turbo.png
- text: >-
A deer grazing in the forest surrounded by dense trees and the sun bright
in the sky. ethereal motion blur style
output:
url: images/04_turbo.png
- text: >-
A highspeed train going through a mountain tunnel in the distant
background, rice fields in the foreground. Zoomed out shot. ethereal
motion blur style
output:
url: images/05_turbo.png
- text: >-
A snake wrapped around a katana with its eyes closed. ethereal motion blur
style
output:
url: images/06_turbo.png
- text: >-
A photographer holding his camera and leaning out of his jeep on a grassy
plain. ethereal motion blur style
output:
url: images/07_turbo.png
tags:
- lora
Krea 2 LoRA — Sunsetblur
A LoRA for Krea 2 — Sunsetblur (trigger: ethereal motion blur style).
- Trigger word:
ethereal motion blur style - To be used on:
krea/Krea-2-Turbo, the few-step distilled checkpoint shown in the previews above. - Trained on:
krea/Krea-2-Raw. - Weights:
sunsetblur.safetensors - Previews: rendered on Turbo at 8 steps, guidance 0.0, LoRA weight
0.8.
Usage
import torch
from diffusers import Krea2Pipeline
pipe = Krea2Pipeline.from_pretrained("krea/Krea-2-Turbo", torch_dtype=torch.bfloat16).to("cuda")
pipe.transformer.load_lora_adapter("krea/Krea-2-LoRA-sunsetblur", weight_name="sunsetblur.safetensors")
pipe.transformer.set_adapters("default", weights=0.8)
prompt = "A deer grazing in the forest, ethereal motion blur style"
image = pipe(prompt, num_inference_steps=8, guidance_scale=0.0).images[0]
image.save("sunsetblur.png")